1 00:00:04.822 --> 00:00:15.694 I'm going to talk about the current state of aggregate economics. I'm not going to be presenting scientific results. 2 00:00:15.694 --> 00:00:30.319 I'm focusing this discussion on the young economist. This is a great time in aggregate economics. 3 00:00:30.319 --> 00:00:42.069 So much has been learned in recent years, so much is being learned but more important so much remains to be learned. 4 00:00:42.069 --> 00:00:54.633 And when you young economists, as one of you have already, figured things out, I expect you to educate me. 5 00:00:54.633 --> 00:01:03.757 When I use aggregate economics, why I use aggregate economics and not macroeconomics, 6 00:01:03.757 --> 00:01:18.585 well macroeconomics I see as a serious but failed science. It was largely separated from economics in the '50's and '60's. 7 00:01:18.585 --> 00:01:27.648 The core of macroeconomics is the empirically determined dynamic system governing the evolution 8 00:01:27.648 --> 00:01:33.336 of the national income and product statistics. 9 00:01:33.336 --> 00:01:44.211 The view was that the economic foundations of the empirically determined dynamic systems would subsequently be developed. 10 00:01:44.211 --> 00:01:55.585 This was not to be. The Lucas critique, it's inconsistent with the dynamic economic theory. 11 00:01:55.585 --> 00:02:03.463 The existence of policy invariant laws of motions, questions of what should you do now, what will happen 12 00:02:03.463 --> 00:02:14.339 if you do this or that now are not well posed economic questions. 13 00:02:14.339 --> 00:02:21.900 By the way I don't use microeconomics either, there's just economics. 14 00:02:21.900 --> 00:02:35.964 And there's so much interaction between people working in all kinds of different areas of substantive interest. 15 00:02:35.964 --> 00:02:40.835 What we should share is common tools. 16 00:02:40.835 --> 00:02:58.340 And the results in financial and labour and others restrict the models that we use to, I don't know, evaluate a tax policy rule. 17 00:02:58.340 --> 00:03:04.966 It seems like the logical thing to do but it was somewhat radical when Finn Kydland 18 00:03:04.966 --> 00:03:14.965 and I used the aggregate growth model to study business cycles. The growth model is not part of macroeconomics. 19 00:03:14.965 --> 00:03:21.416 It was part of microeconomics, what they called microeconomics in yesteryear. 20 00:03:21.416 --> 00:03:27.727 By the way there is one person who is a little bit away ahead of his time, 21 00:03:27.727 --> 00:03:40.854 they call him the father of modern finance Merton Miller, but he's also using that growth model to study aggregate economics. 22 00:03:40.854 --> 00:03:48.103 He was not happy with the old theory when he tried to teach macro at the university of Chicago business school. 23 00:03:48.103 --> 00:03:59.481 So he wrote a book, I think it's the Neo Classical Growth Model in Macroeconomics or some title close to that with Charles Upton. 24 00:03:59.481 --> 00:04:14.355 What have we learned about business cycles, using the growth model? Well to answer that let's look at some statistics or a plot. 25 00:04:14.355 --> 00:04:21.354 Let's look at the last 50 years of fluctuations in the United States. These are population corrected. 26 00:04:21.354 --> 00:04:31.982 I always use working age population and as a proxy for working age population I use people aged 16 to 64, 27 00:04:31.982 --> 00:04:38.608 it's not a perfect proxy. I always correct for trend. 28 00:04:38.608 --> 00:04:48.233 There's been a constant secular growth of nearly 2% a year in living standards due to growth in productivity. 29 00:04:48.233 --> 00:05:05.608 There's been this accumulation of knowledge. This results in a doubling of living standards every generation. 30 00:05:05.608 --> 00:05:12.168 I tell my kids how much better off they are than I was when I was a kid 31 00:05:12.168 --> 00:05:22.231 and I tell them how much better off their kids are than when they were a kid. 32 00:05:22.231 --> 00:05:33.169 Here you see the fluctuations from the '59 to nearly 2009. 33 00:05:33.169 --> 00:05:43.792 You see fluctuations, why, these are deviations from trend, the flat line is trend growth, 34 00:05:43.792 --> 00:05:58.044 living standards doubling every generation. The big expansion in the '60's, well we now can identify what gave rise to that. 35 00:05:58.044 --> 00:06:11.797 Turned out to be productivity. That was an error, a major innovation in the United States. 36 00:06:11.797 --> 00:06:20.796 Jet airplanes, mainframe computers, the chemical industry, spectacular advances. 37 00:06:20.796 --> 00:06:30.794 And the interstate highway system that dramatically reduced transportation costs. 38 00:06:30.794 --> 00:06:38.358 The longest expansion was in the '80's and we know that that was tax rate cut driven. 39 00:06:38.358 --> 00:06:48.107 And the recovery of productivity growth. Finn and I didn't have taxes in our model but we estimated up to 1980, 40 00:06:48.107 --> 00:07:00.296 so we would have been way off having abstracted from taxes in that period. 41 00:07:00.296 --> 00:07:11.234 The biggest recession, my recessions are definition, the MBR uses the 0% trend 42 00:07:11.234 --> 00:07:19.046 and they used all kinds of judgement and they all use the highly preliminary statistics. 43 00:07:19.046 --> 00:07:31.171 They're done on a real time basis. And the statistics that come in later for GDP are a lot different. 44 00:07:31.171 --> 00:07:44.234 They may find an extra 5 million workers in the censuses of households that are every 10 years or every 5 years for businesses. 45 00:07:44.234 --> 00:08:01.736 And a few years later you see the major revisions. The recent recession in 2008/2009 with the revised statistics of GDP, 46 00:08:01.736 --> 00:08:13.671 incorporating some of the census information and new benchmarks, is not looking so puzzling. 47 00:08:13.671 --> 00:08:23.986 One of the things that surprised me was that, well that it seemed to be that the financial crisis 48 00:08:23.986 --> 00:08:34.362 and monitory factors didn't play much of a role, they play a big role in filling up space in newspapers. 49 00:08:34.362 --> 00:08:42.111 But the quantitative models showing that they have, using them to account, 50 00:08:42.111 --> 00:08:54.924 there's just not much residual in the models that abstract from these features of reality. 51 00:08:54.924 --> 00:09:01.986 One of the major themes that is exciting developments in recent years is the model economies are restricted 52 00:09:01.986 --> 00:09:07.362 by both individual and aggregate observations. 53 00:09:07.362 --> 00:09:15.862 It used to be, well there's just some laws, emotion in using aggregate only and any theorist would say 54 00:09:15.862 --> 00:09:26.923 thou shall not touch a number, that was a taboo in the old days. And they used to say I can explain anything. 55 00:09:26.923 --> 00:09:37.237 By the way using general equilibrium theory you can. Some animal spirits or something of that sort. 56 00:09:37.237 --> 00:09:52.926 That would preference shocks in the Arrow-Debreu state contingent, I should say event contingent on language. 57 00:09:52.926 --> 00:10:03.800 What were Finn and I trying to do was to come up with a propagation mechanism for monitory shocks. 58 00:10:03.800 --> 00:10:17.988 Wicksell and Frisch, the great, I guess one is a Swede, one is a Norwegian, economist, emphasised this propagation. 59 00:10:17.988 --> 00:10:24.614 Finn and I were both out of the operations research tradition. 60 00:10:24.614 --> 00:10:35.926 So we were naturally quantitative and used the computers early when it was not so easy as it is today. 61 00:10:35.926 --> 00:10:39.990 We worried about the numbers, how big things were. 62 00:10:39.990 --> 00:10:45.614 In this time to build, the question is how long did it take to build the new office building, 63 00:10:45.614 --> 00:10:57.613 how big were office buildings as expenditures in this field. 64 00:10:57.613 --> 00:11:06.927 So we put those restrictions on sort of the value put in place in the construction industry. 65 00:11:06.927 --> 00:11:14.866 We insisted that the aggregate assumptions be consistent with individual observations. 66 00:11:14.866 --> 00:11:20.491 These individual observations restricted the aggregate model. 67 00:11:20.491 --> 00:11:31.113 Our basic finding, that if the aggregate household had high labour supply, elasticity, notice I bold faced aggregate. 68 00:11:31.113 --> 00:11:41.679 It's going to be analogous to aggregate production function. Productivity shocks are persistent with the given variants. 69 00:11:41.679 --> 00:11:51.365 Then a predication of the growth model as the economy will display business cycle fluctuations of the nature observed. 70 00:11:51.365 --> 00:11:58.866 By the way time to build turned out to be relatively unimportant as shown by Garry Hanson, 71 00:11:58.866 --> 00:12:14.867 who wrote that beautiful strip down version of the model that captured all the essential features. 72 00:12:14.867 --> 00:12:20.943 The productivity shocks had to be persistent, well we could look at these residuals 73 00:12:20.943 --> 00:12:30.070 to the aggregate production function and they were. And the variant was of the appropriate size. 74 00:12:30.070 --> 00:12:39.632 We said that these factors in this period of I think '54 to '80 accounted for 70% of the variants. 75 00:12:39.632 --> 00:12:48.943 We used variants measure rather than standard deviation because 85 would have sounded too big, percent. 76 00:12:48.943 --> 00:13:03.695 But if factors or orthogonal variances add and if 70% leaves a lot, if there was no productivity shocks, 77 00:13:03.695 --> 00:13:10.821 it would still be as 50% as variable. That's just the way variances work. 78 00:13:10.821 --> 00:13:16.322 People were understandably bothered that the variance was so large. 79 00:13:16.322 --> 00:13:23.320 The growth of the stock of useful knowledge is relatively smooth. 80 00:13:23.320 --> 00:13:37.319 Lots of people are doing this and occasionally there's a small burst of technologies we found but not as big. 81 00:13:37.319 --> 00:13:46.884 And then we see these, we saw these huge differences in productivity across countries, 82 00:13:46.884 --> 00:13:55.196 that accounts for virtually all the difference in living standards over time and across countries. 83 00:13:55.196 --> 00:14:02.508 Productivity is used to residually measure the performance of the regulatory 84 00:14:02.508 --> 00:14:12.444 and legal institutions governing the business sector and not just change in the stock of knowledge. 85 00:14:12.444 --> 00:14:22.259 We have the first source, be country specific, the second source is common. 86 00:14:22.259 --> 00:14:37.073 As I said most difference in, due to differences in productivity. I emphasise this aggregate labour supply elasticity is big. 87 00:14:37.073 --> 00:14:44.697 Prediction of aggregation theory are that the aggregate elasticity will be much greater 88 00:14:44.697 --> 00:14:55.697 than the individual elasticity of the people being aggregated under certain conditions and equal under others. 89 00:14:55.697 --> 00:15:06.508 Which one, if conditions are such that the principle margin of adjustment is the number of working 90 00:15:06.508 --> 00:15:17.981 and not in work week length, the aggregate elasticity is much larger than the individual elasticity's. 91 00:15:17.981 --> 00:15:22.605 This is confirmed by observations. 92 00:15:22.605 --> 00:15:33.542 If they were the same there'd be a puzzle, there'd be a deviation from theory, there would be an inconsistency. 93 00:15:33.542 --> 00:15:42.228 Because empirically the principle margin is the number of working, so both individual and aggregate observations are consistent. 94 00:15:42.228 --> 00:15:51.105 There would be a puzzle if these elasticity's were equal. Other aggregate observations are also relevant. 95 00:15:51.105 --> 00:15:58.167 There used to be only data basis for the US and maybe UK but now it's so easy to access them 96 00:15:58.167 --> 00:16:03.230 for just about any country in the world. 97 00:16:03.230 --> 00:16:14.044 Across countries there's large difference in market hours per working age person. 98 00:16:14.044 --> 00:16:24.043 And in the marginal effective tax rates across countries and also across time within a given country. 99 00:16:24.043 --> 00:16:30.543 These observations give a tight estimate of the aggregate supply elasticity. 100 00:16:30.543 --> 00:16:34.918 This estimate is the same as the one based on the business cycle observations. 101 00:16:34.918 --> 00:16:39.168 Things had to be consistent, you can't have one set of assumptions to nail this correlation 102 00:16:39.168 --> 00:16:43.419 and then you go on and say I've done it. 103 00:16:43.419 --> 00:16:51.354 Then create a different inconsistent set to nail some other observations and say now we understand why this is. 104 00:16:51.354 --> 00:16:58.794 The 2, the assumptions you make have to be, you know restricted by observation and be consistent. 105 00:16:58.794 --> 00:17:02.795 And we got to build, it's a cumulative. 106 00:17:02.795 --> 00:17:15.984 The big current development in aggregate labour supply theory is the determination of the fraction of life time worked. 107 00:17:15.984 --> 00:17:24.733 We heard a lot about retirement, people are getting older, even you young economists, 108 00:17:24.733 --> 00:17:29.420 it's predictable that every year you'll get a year older. You can't insure against that. 109 00:17:29.420 --> 00:17:52.795 It is important for predicting the consequence of pension reform. 110 00:17:52.795 --> 00:18:00.796 Rogerson and Wallenius, and Wallenius done this, Wallenius is here, she has major contribution 111 00:18:00.796 --> 00:18:03.669 and maybe she should be the one up here. 112 00:18:03.669 --> 00:18:17.045 The typo is, the editor must have, it's LEN, not M, she's at the Stockholm school of economics. 113 00:18:17.045 --> 00:18:23.233 Well the macro cross countries, that's a 45 degree line, this goes off the margin rate 114 00:18:23.233 --> 00:18:30.610 as substitution equal the marginal rate of transformation for the aggregate production function 115 00:18:30.610 --> 00:18:52.045 and the aggregate utility flow function. It's not that the French are lazy, it's just that they have high tax rates. 116 00:18:52.045 --> 00:19:06.297 And there's no free parameters here, there's one actually, that's all but that's common to all the countries. 117 00:19:06.297 --> 00:19:17.858 The major business cycle findings, their optional responses to real shocks, taxes and demography and productivity are real. 118 00:19:17.858 --> 00:19:23.797 Attempts to stabilise the economy are misguided and will just add a lot of confusion and noise 119 00:19:23.797 --> 00:19:43.109 and uncertainty which will have cost. By the way this simple model predicts that Finn and I used and Handson simplified, 120 00:19:43.109 --> 00:19:49.862 predicts well until a big puzzle arose in the '90's, in the decade. 121 00:19:49.862 --> 00:19:59.736 What happened in the '90's, when briefing the president of the Minneapolis federal reserve bank 122 00:19:59.736 --> 00:20:09.422 we said why were people working so damn much, did they get a contagious case of workaholism. 123 00:20:09.422 --> 00:20:18.859 There wasn't any cut in taxes, there wasn't any major labour market reforms. 124 00:20:18.859 --> 00:20:29.673 But the productivity was up, that's GDP per hour, per market hour was only up modestly. 125 00:20:29.673 --> 00:20:38.674 Yet ours was way up. In terms of the boom normally 2/3 hours 1/3 productivity. 126 00:20:38.674 --> 00:20:58.736 In that boom 125% hours, 25% productivity. That problem bothered me for 10 years before being resolved. 127 00:20:58.736 --> 00:21:08.549 Ellen McGrattan and I looked at the stock market and found that to really get the fundamentals intangible capital 128 00:21:08.549 --> 00:21:17.050 is big time, that is as big time as tangible capital, the things that show up in the national accounts. 129 00:21:17.050 --> 00:21:21.424 The things that are capitalised. 130 00:21:21.424 --> 00:21:30.301 Intangible capital is R&N, the developing new products, building an organisation. 131 00:21:30.301 --> 00:21:42.925 And one of the most important of them all is investment in the human capital of young people on the job. 132 00:21:42.925 --> 00:21:50.862 Part of that is financed by businesses and that is specific to that business. 133 00:21:50.862 --> 00:22:00.612 That is capital of that business. The other part is, Paul Romer calls it rival human capital. 134 00:22:00.612 --> 00:22:09.799 You can walk down the street with your great skills and get an even higher paying job at another university 135 00:22:09.799 --> 00:22:16.612 or other research centre. 136 00:22:16.612 --> 00:22:25.183 When I say predict and actual close, what do I mean by that, we take productivity, population and tax rates as exogenous, 137 00:22:25.183 --> 00:22:27.497 and initial capital stock. 138 00:22:27.497 --> 00:22:34.370 We construct the equilibrium path under a perfect foresight assumption which is an incorrect assumption. 139 00:22:34.370 --> 00:22:38.743 People don't have perfect foresight. 140 00:22:38.743 --> 00:22:46.872 We of course run some experiments and found out that if we, were we generated data artificially on the computer, 141 00:22:46.872 --> 00:22:56.058 that we knew how people, what the rational expectation, expectation scheme were. 142 00:22:56.058 --> 00:23:04.810 And the question is how different were the path based upon this incorrect assumption and the one based upon the correct one. 143 00:23:04.810 --> 00:23:06.619 We created that economy. 144 00:23:06.619 --> 00:23:13.807 By the way you create your economies, your gods, you have people in your economy and they're artificial 145 00:23:13.807 --> 00:23:24.808 but they're useful for building economic intuition. 146 00:23:24.808 --> 00:23:30.809 We learned that by, and it turned out that the predicted and actual were quite close until we hit the 90's 147 00:23:30.809 --> 00:23:38.249 and then there was a huge deviation from theory. Deviations from theory drives the development of theory. 148 00:23:38.249 --> 00:23:46.871 That's not just, that's true of any hard science which aggregate economics has become. 149 00:23:46.871 --> 00:23:52.874 There is cases where the expectations do matter. 150 00:23:52.874 --> 00:24:07.436 Ellen McGrattan has, it's going to be an important, I expect, paper and using it to the great contraction in 1930, 151 00:24:07.436 --> 00:24:30.125 about 25% and the not so great contraction that we recently experienced of the order of 10%. 152 00:24:30.125 --> 00:24:38.621 Now questions about the 1990 boom, we have to look at why were corporate accounting profits so low. 153 00:24:38.621 --> 00:24:42.560 By the way corporate accounting profits are high now. 154 00:24:42.560 --> 00:24:49.497 Why were they low, everybody knows that it was a boom then? The MBA's were dropping out of business school. 155 00:24:49.497 --> 00:24:59.749 Why? They wanted to become entrepreneurs, start businesses, get rich, create social surplus. 156 00:24:59.749 --> 00:25:11.622 And why were people working so much? I say no tax cut, no labour market reform. And I did say that it bothered me for a decade. 157 00:25:11.622 --> 00:25:16.436 The resolution was the intangible investment introduced, this was important to get of a theory 158 00:25:16.436 --> 00:25:27.562 the fundamental value of the stock market that works so well for the secular movements of the stock market relative to the GDP. 159 00:25:27.562 --> 00:25:33.562 In the UK it varied by a factor of 3 in the 1960 to 2000 period. 160 00:25:33.562 --> 00:25:50.623 Most things, ratios relative to GDP, the variations are small and it was 21/2 times in the US. 161 00:25:50.623 --> 00:25:57.188 When you look at the value of many corporations, Coca-Cola is worth a lot more than its factories, 162 00:25:57.188 --> 00:26:05.312 it's that little formula in the safe in Atlanta, the formula for making Coca-Cola, 163 00:26:05.312 --> 00:26:14.813 the brand name is, but there's also Weaties and Cheerios, General Mills, Microsoft, 164 00:26:14.813 --> 00:26:25.376 Sysco, oh I guess now what's the corporation worth the most, I think I read it was Apple. 165 00:26:25.376 --> 00:26:33.379 They have a lot of technology, no how, brand names that is embodied within the organisation. 166 00:26:33.379 --> 00:26:44.752 And this form of capital is important. It was needed to account for the behaviour of the stock market. 167 00:26:44.752 --> 00:26:48.938 Essentially the value of the stock market equals the value of the product of assets, 168 00:26:48.938 --> 00:27:00.063 tangible and intangible at the cost of these to the owners in terms of their consumption. 169 00:27:00.063 --> 00:27:10.813 Taxes and institutional arrangements matter and matter a lot there. 170 00:27:10.813 --> 00:27:19.315 But we do, we introduce that and we said, well let's let the productivities be different in these 2 activities. 171 00:27:19.315 --> 00:27:36.189 One of producing tangible capital, buildings, machines, etc. and vehicles, another was intangible capital. 172 00:27:36.189 --> 00:27:51.315 We residually determined those 2 productivity series and treated them as exogenous. 173 00:27:51.315 --> 00:27:55.379 There was a wealth of evidence that they were, you sort of see the way they shift the people, 174 00:27:55.379 --> 00:28:02.504 we went down and looked at these household surveys and what occupations people are in, 175 00:28:02.504 --> 00:28:11.126 either in the industries that start, occupations that make major investments in activities 176 00:28:11.126 --> 00:28:19.006 that are part of the investments in this intangible capital. And yes there was a big shift. 177 00:28:19.006 --> 00:28:31.127 And by the way the NSF publishes R&D numbers which is an important component of intangible capital investment, not stock. 178 00:28:31.127 --> 00:28:36.751 And there's a problem of making up a depreciation rate to get up to a stock. 179 00:28:36.751 --> 00:28:41.006 We could only measure net investment. 180 00:28:41.006 --> 00:28:48.190 But it turned out the depreciation rate for our purposes, the results were insensitive to that. 181 00:28:48.190 --> 00:29:02.380 If they were, they were. And we couldn't say, in our method it would be not that useful. 182 00:29:02.380 --> 00:29:08.315 We even looked at the time profile of wholly owned American subsidiaries. 183 00:29:08.315 --> 00:29:15.254 You know US corporations, the ones that are publicly traded, they produce about, schedule C corporations, 184 00:29:15.254 --> 00:29:19.817 they produce half the output, these are the ones that pay corporate income tax. 185 00:29:19.817 --> 00:29:27.005 There's a big component, sizable component of corporations that distribute everything to their owners, 186 00:29:27.005 --> 00:29:34.317 workers and it gets treated as ordinary income for tax purposes. 187 00:29:34.317 --> 00:29:43.318 And they pay no corporate income tax. The schedule S corporations. But now I'm going to talk about the schedule C. 188 00:29:43.318 --> 00:29:47.757 Nearly half of the profits of American corporations according to accounting profits 189 00:29:47.757 --> 00:29:59.194 come from American wholly owned foreign subsidiaries, in 10% of their, a little bit over 10% of their investment. 190 00:29:59.194 --> 00:30:04.632 Any accountant should be worried about and wonder what's going on. 191 00:30:04.632 --> 00:30:13.692 The thing though is this expensive investment reduces your accounting profits. 192 00:30:13.692 --> 00:30:21.758 Intangible investment is not counted as part of output. And that lowers the profit. 193 00:30:21.758 --> 00:30:28.008 And these foreign subsidiaries are not very profitable accounting wise initially 194 00:30:28.008 --> 00:30:35.630 because that's when their big intangible investment is being made. And later on they get profitable. 195 00:30:35.630 --> 00:30:43.130 We didn't think we could get those statistics from the IRS and we couldn't. 196 00:30:43.130 --> 00:30:49.505 It would be easy to carry out the exercise but there's confidentiality problems. 197 00:30:49.505 --> 00:31:00.194 Fortunately we found some people that had looked at these statistics and it was in dramatic conformity. 198 00:31:00.194 --> 00:31:11.631 But where does this thing show up, these capital gains. Capital gains are not part of measured output. 199 00:31:11.631 --> 00:31:18.195 But you can see something about the capital gains by studying the balance sheet, the flow of funds, 200 00:31:18.195 --> 00:31:25.511 statistics have the household balance sheets and we can see the holding period or the private sector, there's foundations, 201 00:31:25.511 --> 00:31:32.570 non profits are with the household sector though there's a small part. 202 00:31:32.570 --> 00:31:44.820 And this holding period gains are just of the right magnitude. Normally about 2% a year of GDP. 203 00:31:44.820 --> 00:31:57.633 In the late '90's they went up to 6% so they were under measuring output by an abnormally low large amount. 204 00:31:57.633 --> 00:32:05.444 By the way in the 1990's we all knew the economy boomed. 205 00:32:05.444 --> 00:32:12.009 Here's sort of the, using the naïve model without intangible capital. 206 00:32:12.009 --> 00:32:19.195 Actually the red line per capita hours, blue line predicted per capita hours. 207 00:32:19.195 --> 00:32:34.775 Big deviation. Put the intangible capital in, things fit almost like a glove. 208 00:32:34.775 --> 00:32:43.212 By the way you could always introduce enough preference parameters to fit any set of observations. 209 00:32:43.212 --> 00:32:53.900 Why do Europeans work 30% less than Americans, at least until recently and in other advanced industrial countries? 210 00:32:53.900 --> 00:33:02.149 Some people say western Europeans just get more disutility from work or more utility from vacations, 211 00:33:02.149 --> 00:33:11.840 they're not that different is my working, my maintained hypothesis. 212 00:33:11.840 --> 00:33:21.837 And the Europeans used to work as much as, western Europeans as the people in the other advanced industrial countries. 213 00:33:21.837 --> 00:33:33.212 It's not a contagious case of laziness. Rely on that is, in the animal spirits and like, that's not science. 214 00:33:33.212 --> 00:33:39.088 We have a factor that we can go out there and measure, tax rates. 215 00:33:39.088 --> 00:33:45.963 Europeans have 60% marginal effective tax rates while the rest of the world has 40% and that's enough, 216 00:33:45.963 --> 00:33:51.215 that gives that predicted difference. 217 00:33:51.215 --> 00:34:02.277 Now everybody says financial crisis, financial crisis, that's why the US economy is in the tanks, 218 00:34:02.277 --> 00:34:14.464 I say no, I say the Fed and the treasury and the European central banks and other institutions did a great job 219 00:34:14.464 --> 00:34:26.152 in September of 2008 avoiding a liquidity crisis. 220 00:34:26.152 --> 00:34:37.215 There used to be regular liquidity crisis prior to the formation of the Federal Reserve Bank in 1913, every 4 or 5 years, 221 00:34:37.215 --> 00:34:52.651 every 5 years about. The institutions we set up did not serve us well then. 222 00:34:52.651 --> 00:35:07.780 There would have been some real disruption to the economy. People tried desperately to come up with the model of this. 223 00:35:07.780 --> 00:35:15.465 That is sort of consistent with individual and aggregate observation, they failed. 224 00:35:15.465 --> 00:35:23.279 And these people are good people so it's not, lack of talent it is not the reason. 225 00:35:23.279 --> 00:35:38.965 By the way there's been, using this model we looked at the great depressions, Cole and Ohanian initiated this in about, 226 00:35:38.965 --> 00:35:51.779 very late 1990's. They broke a taboo and looked at the great depression using this neoclassical growth model. 227 00:35:51.779 --> 00:36:03.905 Using the same methodology we systemically, a number of people, examined 16 great depressions. 228 00:36:03.905 --> 00:36:14.590 The US current depression and the Italy's depression in the 1930's were not so great depressions, 229 00:36:14.590 --> 00:36:27.467 they didn't quite hit that big magnitude that we insisted for great. These depressions are not a thing of the past. 230 00:36:27.467 --> 00:36:33.719 And I could list a number of, Switzerland had their great depression and New Zealand in 1970 to 1995. 231 00:36:33.719 --> 00:36:46.906 Chile and Mexico beginning in the 1980's, Chile recovered, Mexico didn't. 232 00:36:46.906 --> 00:36:58.970 And there's lots of open puzzles associated with the depression. What happened in that 1930, why did people stop investing. 233 00:36:58.970 --> 00:37:08.283 Some people say they couldn't get money to invest, why did these businesses make these huge distribution to the owners. 234 00:37:08.283 --> 00:37:20.532 Dividends went way, way up. Investment is mostly financed by retained earnings for corporations. 235 00:37:20.532 --> 00:37:27.033 They had access to the money or else they wouldn't distribute it to the owners. 236 00:37:27.033 --> 00:37:31.220 It's a puzzle that's been bothering me for so long. 237 00:37:31.220 --> 00:37:39.219 I think that back in 1976 I heard the LeRoy-Porter paper in Madison Wisconsin, 238 00:37:39.219 --> 00:37:52.908 summery econometric society meetings on this excess volatility. And Shiller independently made the same observation. 239 00:37:52.908 --> 00:38:05.282 That puzzle is still open. And when one of you resolve it, call me up, please. 240 00:38:05.282 --> 00:38:14.220 Well the secular path, I mentioned that already. Another big puzzle, why are there such large gains from openness. 241 00:38:14.220 --> 00:38:24.720 There's been a really exciting development in trade initiated by Eaton and Kortum. 242 00:38:24.720 --> 00:38:33.095 And they're matching up with the micro observations at the firm level. 243 00:38:33.095 --> 00:38:38.283 These are in the comparative advantage tradition. 244 00:38:38.283 --> 00:38:49.095 There is also the increasing returns of Dixit-Stiglitz model extended by Krugman to international trade. 245 00:38:49.095 --> 00:38:52.658 And also the classical one. 246 00:38:52.658 --> 00:38:59.721 By the way all these models, when people were restricted to observation, observed flows, give way too small gains from trade. 247 00:38:59.721 --> 00:39:10.096 I like this figure, these are Madison numbers, the GGDC Geary-Khamis purchasing power parity. 248 00:39:10.096 --> 00:39:26.408 EU productivity relative to the US, I use productivity because at the later part hours in Europe fell relatively. 249 00:39:26.408 --> 00:39:35.911 The Europeans are, western Europeans are, used to only be about 55% as productive as the Americans for about 50 years, 250 00:39:35.911 --> 00:39:39.035 prior to world war 11. 251 00:39:39.035 --> 00:39:50.162 Since 1957 Treaty of Rome, the original EU 6 caught up and has stayed up there in terms of productivity. 252 00:39:50.162 --> 00:39:58.472 Why? The late joiners, my friends from Poland and this is one of the important, 253 00:39:58.472 --> 00:40:09.159 the biggest country in this set and met some of you people from Estonia at dinner last night and had fascinating discussion. 254 00:40:09.159 --> 00:40:23.413 There's rapid catch up occurring. This is GDP per capita. In Western Europe it's about 70%. 255 00:40:23.413 --> 00:40:32.661 In Japan as well. But the predicted gains from openness are far smaller. 256 00:40:32.661 --> 00:40:46.975 You know you see this big gain, your theory predicts this amount, what do you do, you scratch your head and become bald. 257 00:40:46.975 --> 00:41:04.771 If not trade then what. Well Ellen McGrattan and I extended the growth model to, well Schumpeter said 258 00:41:04.771 --> 00:41:12.958 you need it for endogenising technology development, you need a monopoly rents. 259 00:41:12.958 --> 00:41:17.710 And that's why everybody goes at monopolistic competition model. 260 00:41:17.710 --> 00:41:23.209 We developed a different way to do it, that stays within competitive equilibrium, 261 00:41:23.209 --> 00:41:31.960 the convex cones that we use in macroeconomics that all the growth accounting builds on, 262 00:41:31.960 --> 00:41:39.895 it's so easy to interact with the data, the extensions. We introduced the concept of a location. 263 00:41:39.895 --> 00:41:53.334 And there's location rents. And we add the composite output. We got a lot bigger gains but it's still way too small. 264 00:41:53.334 --> 00:42:01.085 So that did not. It seemed to be that the diffusion of knowledge is so important and people are trying like mad using that 265 00:42:01.085 --> 00:42:09.396 and modelling that in a productive way. 266 00:42:09.396 --> 00:42:21.648 Another thing is competition. Empirically competition is associated with higher productivity. 267 00:42:21.648 --> 00:42:33.835 Our iron mines up northern Minnesota, productivity doubled in 1982 without any new technology, just change in work rules. 268 00:42:33.835 --> 00:42:44.837 Why did that happen, the Brazilian mines came on line and just that threat of competition was sufficient 269 00:42:44.837 --> 00:42:55.402 to increase the productivity of the mines in northern Minnesota and in Canada and in Sweden for that matter. 270 00:42:55.402 --> 00:43:07.461 It didn't happen in Australia until 10 years later, that's when the Indian iron ore mines came on line and provided, 271 00:43:07.461 --> 00:43:12.835 well this is empirical, it's not theoretical. 272 00:43:12.835 --> 00:43:24.588 And having some better theory there of why these decentralised mechanisms empirically work so much better. 273 00:43:24.588 --> 00:43:38.398 So I conclude. My charge to the next generation of economists, you, make our science an even better one. 274 00:43:38.398 --> 00:43:49.153 I know you will and I think you'll have fun doing this, it's a good profession that we're in, society treats us great. 275 00:43:49.153 --> 00:43:55.774 We live longer, we don't have to be in those tense organisations and we get the chance to, 276 00:43:55.774 --> 00:44:05.774 as soon as the students get beyond their qualifying exams, the first year, they're colleagues. 277 00:44:05.774 --> 00:44:16.338 And I learn so much from them. Any course I teach I tell the students I want to learn more than you learn. 278 00:44:16.338 --> 00:44:23.963 It's just a good time that you've chose to go into this field. 279 00:44:23.963 --> 00:44:33.904 And the little financial crisis have been, the economic crisis have been good, a lot of good people are going into economics now. 280 00:44:33.904 --> 00:44:38.401 And you're the lifeblood of an economics department. 281 00:44:38.401 --> 00:44:39.401 Thank you.